Back to results
US 8,949,233 B2Granted
Adaptive knowledge platform
AssigneeAlexandria Investment Research and Technology, Inc.
InventorsRuey-Lung Hsiao, Eugene Shirley
Life of this patent
Estimating the term…
Abstract
Methods, systems, and apparatus, including medium-encoded computer program products, for providing an adaptive knowledge platform. In one or more aspects, a system can include a knowledge management component to acquire, classify and disseminate information of a dataset; a human-computer interaction component to visualize multiple perspectives of the dataset and to model user interactions with the multiple perspectives; and an adaptivity component to modify one or more of the multiple perspectives of the dataset based on a user-interaction model.
Claims20
A system comprising:
- a knowledge management component to acquire, classify and disseminate information of a dataset;
- a human-computer interaction component to visualize multiple perspectives of the dataset and to model user interactions with the multiple perspectives;
- and
- an adaptivity component to modify one or more of the multiple perspectives of the dataset based on a user-interaction model;
- wherein the human-computer interaction component comprises at least one hardware processor and memory configured to:
- (I) place data elements from a dataset, associated with separate ontologies, into an N dimensional information space based on ontological structure and semantic similarity among ontology elements, and (II) reduce the N dimensional information space based on a received perspective definition for visualization;
- wherein the placement of the data elements comprises, for a given ontology of the separate ontologies and a given parent node Cp in the given ontology with child nodes {C1, C2, . . . , Cx}, and letting S={S1, S2, . . . , Sx, Sp} where Si denotes documents that are assigned to concept Ci:
- (I) calculate term importance values for each term in the documents S to generate a set T={T1, T2, . . . , Tx, Tp} where Ti denotes a set of K terms that have highest term importance values, (II) issue search queries to retrieve results R={R1, R2, . . . , Rx, Rp} where Ri denotes documents returned from the search query with terms Ti, (III) estimate co-relation between (Cp, Ci), where i=1 . . . x, according to Cover(Cp, Ci)=|Rp∩Ri∥Rp|≈P[Ci|Cp], where Cover is the conditional probability of Ci given Cp, and (IV) calculate distance between two concepts Cx and Cy based on Sim(Cx, Cy)=2*P/Px+Py+2*Pr, where Cr is the least common super-concept between Cx and Cy, Px is the length of path from Cx to Cr, Py is the length of path from Cy to Cr, and Pr is the length of path from Cr to the root node;
- and
- wherein the reduction of the N dimensional information space comprises projecting into a three dimensional vector space, having bases α1, α2, α3, in accordance with α1=T1({β1, β2, . . . , βN})=β1, α2=T2({β1, β2, . . . , βN})=β2, and α3=T3({β1, β2, . . . , βN})=β3, where T1, T2, T3 are each a linear or nonlinear combination of the N bases in the N dimensional information space, and β1, β2, β3 are bases for the three retained dimensions.
7 more elements
A computer-implemented method comprising:
- receiving, at one or more computers comprising at least one processor and at least one memory, a definition of a perspective into a dataset associated with N separate ontologies that each specify relationships among data elements in the dataset, wherein N is greater than two, the N separate ontologies correspond to an N dimensional information space, and the data elements are placed in the N dimensional information space according to semantic similarity among ontology elements, wherein the number N of dimensions in the information space matches up with the number N of the separate ontologies;
- reducing, at the one or more computers, the N dimensional information space to X dimensions based on the perspective definition, wherein X is greater than one and less than the perspective comprises a projection from the N-dimensional space to the X-dimensional space, and the perspective definition defines a mapping from original N axes to target X axes, the perspective definition being denotable as PN→3=(F1(DS1), F2(DS2), F3(DS3)), where PN→3 represents a perspective that maps information from N-dimensional space to 3-D space, Fi represents mapping functions from original N-dimensional axes to a target i-th axis, and DSi represents original axes that participate in a transformation;
- and
- presenting a visual representation of the dataset in the X dimensional space.
1 more element
A non-transitory computer-readable medium encoding a computer program product configured to cause data processing apparatus to perform operations comprising:
- receiving a definition of a perspective into a dataset associated with N separate ontologies that each specify relationships among data elements in the dataset, wherein N is greater than two, the N separate ontologies correspond to an N dimensional information space, and the data elements are placed in the N dimensional information space according to semantic similarity among ontology elements, wherein the number N of dimensions in the information space matches up with the number N of the separate ontologies;
- reducing the N dimensional information space to X dimensions based on the perspective definition, wherein X is greater than one and less than N, the perspective comprises a projection from the N-dimensional space to the X-dimensional space, and the perspective definition defines a mapping from original N axes to target X axes, the perspective definition being denotable as PN→3=(F1(DS1), F2(DS2), F3(DS3)) where PN→3 represents a perspective that maps information from N-dimensional space to 3-D space, Fi represents mapping functions from original N-dimensional axes to a target i-th axis, and DSi represents original axes that participate in a transformation;
- and
- presenting a visual representation of the dataset in the X dimensional space.
1 more element
A system comprising:
- a user interface device;
- and
- one or more computers comprising at least one processor and at least one memory configured to interact with the user interface device and to perform operations comprising:
- receiving a definition of a perspective into a dataset associated with N separate ontologies that each specify relationships among data elements in the dataset, wherein N is greater than two, the N separate ontologies correspond to an N dimensional information space, and the data elements are placed in the N dimensional information space according to semantic similarity among ontology elements, wherein the number N of dimensions in the information space matches up with the number N of the separate ontologies;
- reducing the N dimensional information space to X dimensions based on the perspective definition, wherein X is greater than one and less than N, the perspective comprises a projection from the N-dimensional space to the X-dimensional space, and the perspective definition defines a mapping from original N axes to target X axes, the perspective definition being denotable as PN→3=(F1(DS1), F2(DS2), F3(DS3)), where PN→3 represents a perspective that maps information from N-dimensional space to 3-D space, Fi represents mapping functions from original N-dimensional axes to a target i-th axis, and DSi represents original axes that participate in a transformation;
- and
- presenting a visual representation of the dataset in the X dimensional space.
4 more elements
Dependencies are inferred from the claim text.